Coronavirus epidemic: prediction and controlling measures

The COVID-19 outbreak has caused over 1.7 million (still increasing) confirmed cases globally as of April 10th, 2020. The levels of spread and severity of the virus lead to a wide-spread political and economic turmoil. We believe that two critical contributing factors need to be taken into account by the authorities to make effective decisions for controlling the spread of the virus: (i) being familiar with the most effective controlling measures and (ii) having a mathematical model to predict the spread of the virus. In this study, we provided information regarding both of these crucial factors. First, we investigated the importance of different measures such as quarantine, isolation, face mask, social distancing, etc. in controlling the virus in various countries. We then present a mathematical model to predict the spread of the virus in different countries. Our prediction shows an excellent match with the actual data up to now.


Introduction
At the end of 2019, the COVID-19 outbreak originated from Wuhan, a city in central China. The outbreak leads to confinement of over 1 billion people to their homes since the end of January 2020 and indeed tremendously disrupts the health care, economy, and well being. After a while, as the circumstances in The main governing equations of this model are as follows: where t is time, N is the total population size (N = S(t) + I(t) + R(t) = const.), β ∈ {R > 0} is the infection rate, and γ ∈ {R > 0} is the recovery rate.The initial conditions of the governing equations (1-3) are S(0) = S 0 > 0, I(0) = I 0 ≥ 0, and R(0) = 0, respectively.
Using the following minimization problem, we can find the unknown param-75 eters β, and γ to fit the actual data.
where β and γ are design parameters, f (t) = I actual + R actual is the summation of cumulative numbers of the actual number of infected people and recovered ones up to time t, and g(t, β, γ) = I(t) + R(t) is the summation of the predicted values for cumulative numbers of infected people and recovered ones until time t.

80
Note that the design parameters of this minimization problem are the unknowns β and γ.

4
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . Now that we have the values of β and γ, we can use the SIR model to predict the COVID-19 spread behavior in the future. To do so, we introduce additional equations with respect to the expected behavior when the time approaches in-85 finity (t → ∞). We know that as the time approaches infinity, I(∞) = 0, thus R(∞) can be given by Using (3) and (1), S(∞) is By plugging in (6) in (5), we obtain R(∞) as which is a nonlinear equation. By solving this nonlinear equation, we can readily 90 predict the behavior of S(t), I(t), and R(t) 3. Importance of protective measures We believe that the number of individuals who lost their lives due to COVID-19 can be count as a more trustable measure with a lower uncertainty in com-

105
parison with the number of confirmed cases to compare the spread of the virus 5 . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04. 11.20062125 doi: medRxiv preprint in different countries with each other. Thus, in Fig. 1  CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.11.20062125 doi: medRxiv preprint COVID-19 has several features which make the prevention process quite difficult such as the infectivity even before the start of symptoms in the incubation period, relatively long incubation period, transmission from asymptomatic indi-7 . CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.11.20062125 doi: medRxiv preprint viduals, tropism for mucosal surfaces, long duration of the illness, and also the possibility of transmission even after recovery [1, 11,12]. Some provinces in China have enforced compulsory face mask policies in public; however, China's national guideline has adopted a risk-based approach in offering recommendations for using face masks. There is a consistency in the rec-155 ommendation that symptomatic people and those in health-care settings should use face masks, but discrepancies are observed in the community settings. We believe one of the main reasons that some of the countries would not recommend the entire population of the country to use face masks is the shortage in the number of available masks. They would give priority to healthcare workers.

160
For example, the US Surgeon General advised against buying and using masks by healthy people [18]. It makes sense to some extent since the highest risk in COVID-19 is transmission to healthcare workers. In the SARS outbreak of 2002, 21% of the infected ones were healthcare workers [19]. To the best of our knowledge, there is not any clear evidence that face masks can provide effective Noteworthy that individuals who have hypertension, diabetes, COPD, cardiovascular disease, and cerebrovascular disease are at a higher risk. Authorities need to educate the public about the risk of getting COVID-19 in these patients. 180 Furthermore, educating the public in terms of practicing social distancing, washing hands regularly, avoiding touching the face, avoiding travel, etc. can be quite helpful in controlling the spread of the virus.

Prediction
We applied the proposed mathematical model to the data provided by Up to now (April, 10th, 2020), our prediction shows an excellent match with the actual data.

195
It is worth noting that for the countries that the slope of the curve is getting negative, it is quite important to maintain the measures that they considered to control the outbreak. This is due to the fact that the second peak might happen in the situation that the authorities in these countries would not maintain the controlling measures anymore. To prevent the occurrence of the second peak 200 especially in the countries that enforced quarantine measures, it is crucial to release the quarantine order in an appropriate time. They should consider the slope of the increase in the number of daily confirmed cases, the potential of the health-care system, and the number of daily screening tests.

10
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

11
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.
is the (which was not peer-reviewed) The copyright holder for this preprint . https://doi.org/10.1101/2020.04.11.20062125 doi: medRxiv preprint Our prediction shows an excellent match with the actual data up to now.

13
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

15
. CC-BY 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.